Live data
Assembling the frontier
Ranking the latest models and provider endpoints.
Loading model dataLive data
Ranking the latest models and provider endpoints.
Loading model dataThe definitive ranking of open-weight AI models you can self-host, fine-tune, and deploy without restrictions. This page now uses the same live dataset and open-source filtering logic as the Explore leaderboard, so the ordering matches the table view across the app.
| Rank | Model | Quality | Best Price | Top Speed | Max Context | Providers |
|---|---|---|---|---|---|---|
| 1 | Muse Spark 1.3 Meta | 48.1 | $2.00/M | 217 tok/s | 1M | 1 |
| 2 | MiMo-V2.6-Pro Xiaomi | 46.3 | $0.54/M | 319 tok/s | 1M | 3 |
| 3 | Qwen3.8 Max Alibaba | 45.4 | $3.00/M | 37 tok/s | 984K | 1 |
| 4 | Muse Spark 1.3 (xhigh) Meta | 45.1 | $2.00/M | 241 tok/s | 1M | 1 |
| 5 | GLM-5.3 Z AI | 44.8 | $1.70/M | 420 tok/s | 1M | 21 |
| 6 | Kimi K3 Kimi | 43.6 | $4.80/M | 317 tok/s | 1M | 17 |
| 7 | GLM 5.3 Flash Z AI | 41.8 | $0.12/M | 281 tok/s | 1M | 19 |
| 8 | Qwen3.8 2.4T A95B Alibaba | 39.9 | $2.85/M | 178 tok/s | 1M | 7 |
| 9 | Qwen3.8-Flash-Next Alibaba | 39.8 | $0.23/M | 54 tok/s | 1M | 1 |
| 10 | DeepSeek V4.1 Flash DeepSeek | 39.5 | $0.18/M | 598 tok/s | 1M | 21 |
| 11 | DeepSeek V4 Pro 0813 DeepSeek | 36 | $0.88/M | 262 tok/s | 1M | 11 |
| 12 | DeepSeek V4 Flash Vision DeepSeek | 34.8 | $0.32/M | 228 tok/s | 1M | 5 |
Ollama makes it easy to run open-weight models locally. Here are the top picks by hardware tier:
RTX 3070 · M2 MacBook Air
RTX 3090/4090 · M2 Pro/Max
2× RTX 4090 · Mac Studio M2 Ultra
Tip: artifact size is not total runtime memory. Leave room for the context cache, serving framework, operating system, and other applications. The official Qwen3-Coder 30B Ollama artifact is about 19GB; Qwen3-Coder-Next Q4 is about 52GB.
Only models with openly available weights are included. This page uses the same grouped Explore dataset and open-source filter as the live leaderboard, then orders those model families by Artificial Analysis Quality Index. We still surface coding benchmarks here when they are available:
Comprehensive knowledge benchmark across 14 domains. Tests breadth of model capability.
Competition math — tests advanced reasoning. Best signal for math and science tasks.
Contamination-free code generation. Best signal for software development capability.
See live pricing from self-hosting providers, latency, and full benchmark scores for all open source models.
Muse Spark 1.3 leads the live open-source ranking with a Quality Index of 48.1. For price-sensitive deployments, GLM 5.3 Flash is the current budget leader at $0.12/M.
For Ollama and other self-hosted setups, start from the strongest live open-weight models here, then narrow by hardware tier. Muse Spark 1.3 is the best option here when context size matters most.
For most tasks, yes. The top open source models in 2026 trail proprietary leaders by only a small Quality Index margin. The main gaps remain in instruction-following polish, multimodal capability, and very long contexts.
As a rule of thumb, 8GB VRAM is enough for smaller 7B to 8B models, 24GB VRAM is a more practical floor for 30B-class models, and 40GB+ is usually required once you move into 70B territory unless you quantize aggressively. Apple Silicon is also viable for smaller and mid-sized open-weight models when unified memory is high enough.
Muse Spark 1.3 is the strongest coding-oriented open source model on this page. For cheap API access, GLM 5.3 Flash is the most economical current pick. See the full coding LLM ranking for more detail.
All models ranked
🧮AIME 2025 rankings
🖥️Self-hosted picks
🦙Ollama-first picks
📄Best for documents
⚡Compare any 2–4 models
Data sources: Rankings use the same live Explore dataset and open-source filter as the methodology page describes. Explore the filtered table in our interactive leaderboard or compare models side by side.